Skip to main content

Integrations

Info
Last Updated: 2026-01-08

Official and community integrations for Cortex Memory.

Production Ready Integrations

Vercel AI SDK (Production Ready)

Package: @cortexmemory/vercel-ai-provider | Version: v0.27.2 | SDK: v0.21.0+

Add persistent memory to Next.js applications with full streaming support, layer visualization, and memory orchestration.

Features:

  • Automatic memory retrieval and storage
  • Streaming with progressive storage
  • Memory Spaces for multi-tenancy
  • Hive Mode for cross-application memory
  • Interactive quickstart demo
  • Layer flow visualization
  • Edge runtime compatible
  • TypeScript native
  • Self-hosted with Convex

Quick Start:

Option 1: Run quickstart demo (recommended)
$ 

cd packages/vercel-ai-provider/quickstart npm install && npm run dev

Option 2: Use CLI to scaffold project
$ 

cortex init my-app --template vercel-ai-quickstart

Example:

import { createCortexMemory } from "@cortexmemory/vercel-ai-provider";
import { openai } from "@ai-sdk/openai";
import { streamText } from "ai";

const cortexMemory = createCortexMemory({
convexUrl: process.env.CONVEX_URL!,
memorySpaceId: "my-chatbot",
userId: "user-123",
agentId: "my-assistant", // Required in v0.17.0+
});

const result = await streamText({
model: cortexMemory(openai("gpt-4o-mini")),
messages,
});

Learn More:


Integration Guides

Authentication

Status: DIY Integration Guide
Documentation: Auth Providers Guide

Cortex provides a framework-agnostic AuthContext API. This guide shows patterns for integrating with Auth0, Clerk, NextAuth, Firebase, or custom JWT systems.

Warning

Cortex does NOT include specific auth provider packages. You integrate your existing auth system using the AuthContext API.

What you get:

  • Generic integration patterns
  • AuthContext API documentation
  • Example adapters for common providers
  • DIY implementation guidance

What Cortex provides:

  • createAuthContext() function
  • Automatic user/tenant field injection
  • Multi-tenant isolation support
  • Session tracking integration

What you provide:

  • Your auth system (Auth0, Clerk, etc.)
  • Code to extract userId/tenantId from tokens
  • User validation logic

View Auth Integration Patterns →


Planned Integrations

These integrations are planned but not yet implemented:

LangChain.js

Status: Planned | Target: Q2 2026

LlamaIndex.TS

Status: Planned | Target: Q2 2026

MCP Servers

Status: Planned | Target: Q1 2026


Community Integrations

None yet. Build an integration and share it with the community!


Integration Requests

Want an integration with your favorite framework?


Next Steps